© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 49152
    9. 8-RES_FS49152-JB
    11. 26665
    Prev
    Next
    INDEX
    Explanations

    references to a specific pronoun for individuals, particularly focusing on their actions and statements

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    96
    +0.12
    0.6%
    551
    +0.11
    0.5%
    719
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    586
    +0.12
    0.04
    719
    +0.11
    0.04
    139
    +0.11
    0.04
    Negative Logits
    noon
    -0.75
    acters
    -0.69
    rocket
    -0.69
    iencies
    -0.65
    NAT
    -0.63
     disabling
    -0.62
     menstrual
    -0.62
     Measure
    -0.61
    berra
    -0.59
    atible
    -0.59
    POSITIVE LOGITS
     said
    1.18
     replied
    1.16
     wrote
    1.15
     exclaimed
    1.14
     joked
    1.09
     laughed
    1.04
     remarked
    1.03
     says
    1.03
     tweeted
    1.02
    said
    1.02
    Activations Density 0.044%

    No Known Activations